Split "utility" functions from core functionality

[jagerber48]

A few ideas have been identified that would be useful helper functions we could consider adding to uncertainties. It may be appropriate to collect these helpers into a single module, separated out from the core uncertainties code. The following functions are candidates for inclusion in this helpers module:

• covariance_matrix - calculates the covariance matrix between a sequence of UFloat (could be renamed covariance_matrix_from_ufloats
• correlation_matrix - calculates the correlation matrix between a sequence of UFloat (could be renamed correlation_matrix_from_ufloats
• correlated_values - constructs a sequence of UFloat from a sequence of nominal values and a covariance matrix (should be renamed something like ufloats_from_covariance_matrix)
• correlated_values_norm - constructs a sequence of UFloat from a sequence of tuples of values/standard deviations together with a correlation matrix. (Should be renamed something like ufloats_from_correlation_matrix
• ufloat_from_sample_std_dev as discussed in Added ufloat_from_sample function #277
• ufloat_from_sample_sem as discussed in Added ufloat_from_sample function #277
• inverse_variance_weighted_mean as discussed in unumpy.average: init #265

Thoughts? Other candidates for this module? Also what should this module be called? uncertainties.utils or uncertainties.utilities are obvious candidates but there are admonishments like https://breadcrumbscollector.tech/stop-naming-your-python-modules-utils/ to avoid naming modules that way. Honestly, despite the admonishments the utils.name seems fine to me in this case unless there's a better idea.